Molecular genetic diagnosis of kidney ciliopathies: Lessons from interpreting genomic sequencing data and the requirement for accurate phenotypic data

Introduction: Massively parallel sequencing (MPS) techniques have made a major impact on the identification of the genetic basis of inherited kidney diseases such as the ciliopathy autosomal dominant polycystic kidney disease (ADPKD). Great care must be taken when analysing MPS data in isolation from accurate phenotypic information, as this can cause misdiagnosis. Methods: Here, we describe a family trio, recruited to the Genomics England 100,000 Genomes Project, labelled as having cystic kidney disease, who were genetically unsolved following routine data analysis pipelines. We performed a bespoke reanalysis of Whole Genome Sequencing (WGS) data and coupled this with revised phenotypic data and targeted PCR and Sanger sequencing to provide a precise molecular genetic diagnosis. Results: We detected a heterozygous PKD1 frameshift variant within the WGS data which segregated with the redefined ADPKD phenotypes. An additional heterozygous exon deletion in ALG8 was also found in affected and unaffected individuals, but its precise clinical significance remains unclear. Conclusion: This case illustrates that reanalysis of WGS data in unsolved cases of cystic kidney disease is valuable. Clinical phenotypes must be reassessed as these may have been incorrectly recorded and evolve over time. Undertaking additional studies including genotype‐phenotype correlation in wider family members provides useful diagnostic information.


INTRODUCTION
As massively parallel sequencing (MPS) techniques become more accessible for the diagnosis of inherited diseases, the number of genes attributed to a specific phenotype is increasing.As such, correct interpretation of the large volume data outputs produced by techniques such as whole genome sequencing (WGS) is vital to determine the relevant genotypes in families with consistent or overlapping disease phenotypes.Genomic data outputs can also create difficulties when determining the relative phenotypic influence of specific variants in a family, when multiple potentially pathogenic variants are simultaneously discovered (Guaragna et al., 2020).This can fall under one of three differing situations, further complicating the interpretation of the data.The first situation is that of digenic inheritance, where two heterozygous alleles on related disease genes can be a major influence on disease phenotype and is exemplified in diseases such as Alport syndrome (Mencarelli et al., 2015).The second scenario is that of dual molecular diagnosis, where diagnosis following WGS involves two or more discrete clinical diagnoses and two or more discrete disease loci (Posey et al., 2017).This has been described in cases where the different diagnoses have overlapping phenotypic features, and contrarily where the phenotypes have no overlapping features (Posey et al., 2017).The complexity of the potential interaction, or lack of interaction of distinct clinical diagnoses in a patient can complicate the process of reaching an accurate molecular diagnosis (Posey et al., 2017).Morgan et al. highlight the complexity of dual molecular diagnosis in cases of hearing loss, concluding that cases of dual molecular diagnosis can be responsible for up to 10% of genetically unsolved cases, and a deep understanding of phenotype is vital for reaching an accurate molecular genetic diagnosis (Morgan et al., 2022).The final scenario is that of rare genetic variants acting as a phenotypic modifiers.For example, Guaranga et al. describe a case of a family with nephrotic syndrome where the proband was homozygous for a rare MYO1E nonsense variant, while also carrying rare heterozygous alleles in 6 other nephrotic syndrome-associated genes which segregated variably within the family and may have been modifying disease phenotypes (Guaragna et al., 2020).Additional alleles may also contribute directly or indirectly and modify the phenotype of a monogenic disorder, leading to phenotypic variability.Such cases of mutational load have been described in Bardet-Biedl syndrome (Perea-Romero et al., 2022).However, the activity of a modifier allele is much more difficult to identify, and prove to be significant than that of digenic inheritance or dual molecular diagnosis.As well as autosomal recessive disorders, multiple alleles can be identified in autosomal dominant disorders, which again may cause diagnostic confusion (Schrauwen et al., 2018).WGS can also lead to incidental findings of clinical significance (Hegde et al., 2015), which may reveal novel diagnoses, where phenotypes have previously been overlooked.
Autosomal dominant polycystic kidney disease (ADPKD), a kidney ciliopathy, is the most common genetic kidney disease worldwide and is typically caused by pathogenic variants in PKD1 (74-85% of cases (Mantovani et al., 2020)) or PKD2 (15-26% of cases) (Besse et al., 2017).PKD1 and PKD2 encode the polycystin 1 and 2 (PC1, PC2) proteins which are localised to primary cilia (Besse et al., 2017).PC1 is a G-coupled receptor, and PC2 is a calcium (Ca 2+ ) ion channel (Kim et al., 2016).PC1 and PC2 together form a complex that WNT proteins bind to causing a Ca 2+ influx into cells via the PC2 channel in order to regulate tubulogenesis (Kim et al., 2016;Qian et al., 2002).Variation to either PKD1 or PKD2 can therefore lead to perturbed tubulogenesis, which can lead to the formation of kidney cysts (Qian et al., 2002).The main clinical features of ADPKD are enlarged kidneys and multiple bilateral cysts of varying sizes, which can cause pain and secondary infections .Typically, liver cysts are also seen (Besse et al., 2017;Zerres et al., 1985) and these are the most common extra-renal manifestations (Goksu & Khattar, 2022).Overall, the PKD1 phenotype is more severe than the PKD2 phenotype.The average age of onset of kidney failure for all PKD1 variants is 58.1 years, while for PKD2, it is 79.7 years (Harris & Torres, 2002).
Within PKD1, truncating variants have a more severe phenotype than non-truncating variants, with kidney failure occurring at an average age of 55.6 years in those with truncating PKD1 variants, and 67.9 years in those with non-truncating PKD1 variants (Cornec-Le Gall et al., 2013;Harris & Torres, 2002).Cases of biallelic PKD1 (Al-Hamed et al., 2019) and PKD2 variants (Durkie et al., 2021) have been reported leading to earlier and more severe kidney disease.In addition, patients with combinations of both PKD1 and PKD2 alleles have been reported (Durkie et al., 2021).In parallel, investigation into milder cystic kidney disease phenotypes has been enhanced by WGS approaches.
An increasing number of genes are becoming implicated in PKD.These patients may have an atypical cystic kidney disease phenotype which is less severe than typical ADPKD, and progression to kidney failure is much rarer and or much later in life (Cornec-Le Gall et al., 2018;Shin & Berliner, 2021).Thus, PKD has become increasingly genetically heterogeneous.DNAJB11 has been recently implicated, encoding an endoplasmic reticulum (ER) protein with a role in maintaining maturation and trafficking of PC1 (Cornec-Le Gall et al., 2018).The kidney disease associated with heterozygous DNAJB11 variants is a combination of ADPKD and autosomal-dominant tubulointerstitial kidney disease, with average sized kidneys, multiple cysts, progressive interstitial fibrosis and late onset kidney failure (Cornec-Le Gall et al., 2018).Another new gene where heterozygous alleles are associated with PKD is IFT140, which encodes part of the IFT-A core complex, responsible for protein trafficking from the cilia tip to the basal body (Senum et al., 2022).The phenotypic features are relatively mild, with large kidney cysts, but minimal kidney insufficiency and hepatic cysts (Senum et al., 2022).Some members of the ALG family of genes, with functions in the (Kukuruzinska & Lennon-Hopkins, 1999) have also been implicated in atypical forms of PKD.ALG5 haploinsufficiency has been associated with a significant decrease in mature PC1, and a fibrocystic phenotype with progressive CKD, and an accelerated disease course in affected individuals once they reach the age of 60 (Lemoine et al., 2022).Heterozygous ALG8 variants are associated with polycystic liver disease, with a mild cystic kidney phenotype.There has been some evidence of an association of heterozygous ALG8 variants with an increased risk of polycystic kidney disease (Apple et al., 2022).ALG8 haploinsufficiency has been linked to a lack of PC1 glycosylation, required for PC1 to traffic to the cilia (Besse et al., 2017).This reduction in PC1 dosage is thought to drive the formation of cysts in both the kidney and liver (Masyuk et al., 2018).There are no reported cases of kidney failure linked to heterozygous, loss of function ALG8 variants (Besse et al., 2019).ALG9 has also been implicated in the atypical kidney disease, with the size and number of kidney cysts milder than in typical ADPKD (Besse et al., 2019).In contrast to ALG8, heterozygous ALG9 variants lead to a predominantly kidney phenotype (Besse et al., 2019).
This increasing genetic complexity, even within a disease phenotype such as cystic kidney disease, increases the challenges of reporting a precise molecular genetic diagnosis and attributing variants in the correct gene leading to the underlying phenotype.The reporting of the major pathogenic allele may be clouded by other alleles of uncertain significance and some uniformity in approaches is desirable (Vears et al., 2017).Here we report a family with a typical ADPKD phenotype in whom WGS analysis had been performed in the context of the Genomics England 100,000 Genomes Project for rare diseases (Caulfield et al., 2019) to determine the underlying genetic cause.Due to incorrect phenotypic data being assigned to a relative, and the identification of more than one potentially pathogenic allele in the proband the family provides useful and valuable learning points in the application of WGS to families with cystic kidney disease phenotypes.

MATERIALS AND METHODS
Participants of the Genomics England 100,000 Genomes Project provided written consent for anonymous use of their clinical and genetic data in research, as approved by the NRES Committee for East England-Cambridge South.

In silico analysis
The 100,000 Genomes Project tiering and exomiser tools were used as a default analysis within the 100,000 Genomes Project.The tiering tool works by annotating variants based on their segregation in the family, frequency in control populations, effect on protein coding, mode of inheritance and whether they are present in the virtual gene panel for the disease (Caulfield et al., 2019).Exomiser filters and prioritises all variants based on likelihood to be causative, by looking at the predicted pathogenicity and allele frequency of the variant in reference databases, and if the patient phenotype matches the expected phenotype for that gene (Caulfield et al., 2019).In addition, we applied MANTAstructural variant and CANVAScopy number variant caller to detect copy number variants (CNVs) in ALG8.A variant-detecting script within Genomics England research environment was used to detect indels in ALG8 and PKD1.Varsome (www.varsome.com)was used to determine pathogenicity ( Kopanos et al., 2019).The Integrated Genomics Viewer (IGV) tool was used to visualise BAM files from WGS data.

Gathering of clinical data and wider family screening
Phenotypic and sequence data for the trio was provided by Genomics England.Additional clinical and diagnostic imaging was obtained from clinical records.

Sanger PCR for PKD1 variant
Sanger sequencing of the PKD1 allele was performed by NHS diagnostic service at Bristol Genetics Laboratory.

CASE REPORT
The proband of the affected family (III: 2) (Figures 1a,b) was a male aged 38 years with typical PKD (with bilateral enlarged polycystic kidneys > 20 cm in length, polycystic liver disease, hypertension, and diverticular disease of the large intestine) (Figure 1c, Table 1) and his two dizygotic twin children were recruited to the Genomics England 100,000 Genomes Rare Disease project for investigation of cystic kidney disease.The proband's daughter (IV:1) (aged 7 years) was reported to have multiple cortical and medullary kidney cysts and was labelled as affected with cystic kidney disease (Figure 1, Table 1).
Her dizygotic twin brother (IV: 2) was also recruited as affected, but no kidney cysts were listed in the phenotypic information for him, suggesting phenotypic uncertainty (Figure 1, Table 1).WGS was performed and the bioinformatics pipeline analysis by Genomics England was reported via the "exit questionnaire" as unsolved, as no suspect or tiered variants were shared between the three individuals.This result in a family with a typical PKD phenotype in the proband and early onset cystic kidney disease phenotype in his daughter prompted us to review the clinical phenotypes in the family, particularly the son whose affected status was in doubt and to perform a bespoke reanalysis of the genomic datasets.
The father of the proband (II:3) had a clinical phenotype consistent with typical ADPKD (Figure 1c, Table 1).He had reached kidney failure at the age of 54 years and had undergone bilateral native kidney nephrectomies (for massively enlarged polycystic kidneys with dimensions of 23 × 11 × 10 cm) and two kidney transplants.He had died aged 66 years from kidney failure.The paternal grandfather (I:2) was also known to have died from kidney failure and cystic kidney disease.
The proband's mother (II:4) was clinically unaffected (Figure 1b).Repeat abdominal ultrasound imaging of the proband's son (IV:2) following an initial abdominal ultrasound which detected suspected cysts, did not identify any evidence cystic kidney disease.Although this does not exclude the future development of kidney cysts, it prompted us to label this family member as unaffected.
As the initial Genomics England analysis returned a negative result, we performed CNV analysis and detected a novel 2509 bp heterozygous ALG8 deletion in both the dizygotic twins (IV:1 and IV:2) and the proband (III:2) (Chr11:g.78105354_78107863del),causing the loss of exon 10 (Figure 2a).We next obtained genomic DNA (extracted  The ALG8 variant is present in II:4, III:2, IV:1 and IV:2.The PKD1 variant is found in III:2, IV:1 and II:3.ACMG classification was provided by Varsome (Kopanos et al., 2019).

Gene
from whole blood) from the proband's parents (II:3 and II:4, Figure 1b).A targeted PCR of ALG8 to identify the deletion confirmed its presence in the probands unaffected mother (II:4), but not the father (II:3) (Figure 2b), leading to uncertainty about the penetrance and pathogenicity of this ALG8 allele.We therefore reviewed the phenotype of family members and redefined the status of IV:2 as unaffected, given a repeat kidney USS was normal, in contrast to his twin sister.Bioinformatic reanalysis of the WGS data on the proband identified another potentially pathogenic allele, a heterozygous known pathogenic PKD1 variant, c.1889del; p.(Pro630Argfs*155) (Zhang et al., 2021) (Table 2).We then performed segregation analysis of the two alleles.Visualisation of the BAM files allowed us to see that the PKD1 allele was present in the proband's affected daughter (IV:1) but not the son (IV:2) (Figure 2c).We deduced that the previous bioinformatic analysis including tiering and exomiser outputs with (IV:2) labeled as affected had excluded the PKD1 allele due to lack of segregation.
Sanger sequencing of II:3 the proband's affected father confirmed that the PKD1 allele had been inherited from the paternal side (Figure 2d).The updated family pedigree shows segregation of the PKD1 allele with the PKD phenotype and a lack of segregation of PKD phenotype with the ALG8 allele (Figure 2e).

DISCUSSION
In the initial trio WGS data analysis carried out by Genomics England, variants found in all three individuals were compared to a cystic kidney disease gene panel, and no causative variant in this trio to explain their diagnosis of cystic kidney disease was detected.A review of the clinical features, an extended family pedigree provided by a genetic counsellor, and bespoke reanalysis of WGS data, including analysis of CNVs, and visual examination of BAM files was performed.
In examining the WGS data from the trio individually, we identified a pathogenic PKD1 variant in individuals III:2 and IV:1, which fitted the phenotypes of these individuals and segregated with known affected family members.This was initially missed by gene panel analysis due to an apparent lack of segregation with the phenotype, due to the mislabeling of IV:2 as affected.Extending the genetic analysis across the available family members confirmed the findings (Table 3).
However, we also identified a rare heterozygous deletion in ALG8, a gene associated with adult-onset cystic kidney and liver disease.This was present together with the PKD1 variant in the proband and the affected daughter.The pathogenicity of this variant is unknown, and our segregation analysis identified this allele in the proband's unaffected mother, suggesting any disease phenotype may not be fully penetrant as has been seen by other ALG8 alleles, where kidney cysts are not always seen (Besse et al., 2017).The ALG8 deletion affects exon 10.To explore if this would be tolerated, gene expression databases were utilized (GTex, 2019).GTEx data suggested that within the examined transcripts, exon 10 of ALG8 is not usually skipped or alternatively spliced, and is expressed in all the tissues, including the kidney (GTex, 2019).In the DECIPHER database there are 12 patients reported to have a heterozygous deletion affecting ALG8, with a variety of phenotypic manifestations.However, none of the patients are reported to have a kidney disease phenotype (Firth et al., 2009).This ALG8 variant remains interesting but remains classed as a variant of uncertain significance (VUS) until further studies can be performed to both determine its pathogenicity alone and its ability to modify the phenotype when it is coinherited with a PKD1 allele as in this case.
ADPKD is a genetically heterogenous disease (Besse et al., 2017;Mantovani et al., 2020;Ravine et al., 1992).Around 10% of cases are listed as genetically unsolved (Audrezet et al., 2012) meaning PKD1 and PKD2 variants have not been identified, or that other gene variants, such TA B L E 3 Key messages from this case.

Key Messages
WGS data must be concurrently analyzed with accurate phenotypic data, and variants confirmed by Sanger sequencing wherever possible to reach an accurate molecular genetic diagnosis Clinical phenotypes must be regularly reassessed to allow for phenotypic evolution and ensure accurate recording of disease phenotypes Wider family segregation studies are invaluable to generate useful diagnostic information Reanalysis of genetically unsolved cases of cystic kidney is valuable given the growing number of genes implicated in ADPKD phenotypes as IFT140 (Senum et al., 2022) are causing the phenotype.As MPS tools such as WGS become increasingly more commonplace in genetic research, an increasing number of genes are being implicated in the pathogenesis of ADPKD.One such gene is ALG8, encoding an α−3glucosyltransferase (ALG8) (Lanktree et al., 2021), which is involved in protein glycosylation (Chantret et al., 2003).
The recessive disease caused by biallelic ALG8 variants is a subtype of Congenital Disorder of Glycosylation (CDG Ih) (Chantret et al., 2003), which causes a severe earlyonset phenotype (Chantret et al., 2003).The phenotypes of heterozygous ALG8 variant carriers are only beginning to be defined.A report published by Besse et al. stated that as of 2019, there were no reported cases of kidney failure caused by heterozygous ALG8 loss of function variants (Besse et al., 2019).Heterozygous ALG8 variants are more commonly associated with autosomal dominant polycystic liver disease (ADPLD), and a less severe kidney phenotype, as many individuals with ALG8 variants have reported multiple liver cysts and almost invariably fewer kidney cysts (Besse et al., 2017).ALG8 and PC1 both interact with GIIβ and SEC63 in the ER protein glycosylation pathway (Besse et al., 2017).ALG8 effects PC1 post-transcriptionally (Besse et al., 2017).Levels of PC1 are reduced in ALG8 negative cells (Besse et al., 2017).As well as this, PC1 is thought to lack glycosylation when ALG8 is not present (Besse et al., 2017).The combined effects of decreased PC1 levels and an abnormal structure due to lack of glycosylation reduces PC1 function (Besse et al., 2017).This reduction in PC1 dosage is thought to drive the formation of cysts in both the kidney and liver (Masyuk et al., 2018).These studies have been carried out in cells with complete ALG8 inactivation, which is different to the situation in heterozygous patients, with at least one fully functional ALG8 allele (Besse et al., 2017).Heterozygous ALG8 loss of function variants therefore cause a decrease in levels of PC1, leading to aberrant cellular signaling, increasing clonal expansion and causing cyst formation (Lanktree et al., 2021).It has been hypothesized that an as-yet undiscovered pathway which relies heavily on negative regulation by PC1/2 could be the overall cause of ADPKD (Ma et al., 2013).Reduced PC1 levels caused by loss of function ALG8 variants could reduce the inactivation of this pathway, leading to the cystic phenotype.Although more commonly associated with ADPLD, a case has been described where a father and daughter carrying the same ALG8 variant have differing phenotypes; the father had ADPLD with a few kidney cysts, while his 19 year old daughter had eight kidney cysts and no hepatic cysts (Besse et al., 2017) suggesting that heterozygous ALG8 variants may encompass both ADPKD and ADPLD diagnoses (Besse et al., 2017).
Early onset cystic kidney disease in individuals with variants in PKD1 and another ADPKD-causing gene has been reported before.A case of a girl with pathogenic variants in both PKD1 and GANAB has been reported with onset at age 12 years, which is an earlier presentation than is typical (Harris & Torres, 2002;Waldrop et al., 2019).There have been some cases of double heterozygous individuals with ADPKD who have variants in both PKD1 and PKD2 (Harris & Torres, 2002).These individuals with digenic disease have a more severe phenotype than monogenic ADPKD (Harris & Torres, 2002).Hence, we also hypothesize that the combined effects of the heterozygous PKD1 and ALG8 alleles could be responsible for more severe phenotypes and early onset compared to typical ADPKD.Further data are required to determine if ALG8 variants can modify the phenotype of ADPKD patients.Typical ADPKD presents clinically in patients over the age of 30 (Fall & Prisant, 2005).The use of kidney ultrasound in at risk individuals does however allow cystic kidneys to be detected at younger ages.In a retrospective cohort study of ultrasound assessment in children and young people under the age of 15 at risk of ADPKD, 193 out of 420 were diagnosed with kidney cysts at a baseline visit (mean age 8.6 +/-4.2 years) (Gabow et al., 1997).The individuals III:2 and IV:1 in the case we describe here have an early age of onset of kidney cysts (12 years and 2 years, respectively) that could indicate an increased risk of severe disease but could also be consistent with a single PKD1 pathogenic variant (Table 1).As IV:2 carries just the ALG8 allele and to date is unaffected by cystic kidney or liver disease, we would predict a very mild, late onset phenotype in this individual.It will be important to follow up the clinical phenotype of IV:2 as they get older to for cystic change within the liver kidneys.

CONCLUSION
This case highlights the need to reanalyse genomic data in a bespoke manner for unsolved cases of cystic kidney disease, and the importance of careful phenotypic annotation.We identified a heterozygous PKD1 frameshift allele which likely explains the phenotype seen in a family with ADPKD, that was unmasked by reanalysis of WGS data.However, we also identified a novel heterozygous deletion variant in ALG8, a gene also implicated in the ciliopathies ADPKD and ADPLD, the clinical significance of which remains unclear.This case highlights the importance of reanalysing WGS data in a bespoke manner for unsolved families and has useful take home messages (Table 3).The PKD1 variant was bioinformatically filtered out in this family due to an apparent lack of segregation with the phenotype, by an automated WGS analysis and variant filtering pipeline.As WGS becomes increasingly more accessible, an integrative technique encompassing both genotypic and phenotypic analysis and perhaps phenotypic reanalysis or review of certain key clinical and biochemical features is needed to ensure accurate conclusions can be to families with inherited diseases.As well as this, uncovering previously hidden variants in WGS datasets coupled with long-term individual phenotypic and clinical follow-up may help to uncover more information about the possible mechanisms and genetic modifiers underlying later onset disorders, such as ADPKD (Brizi et al., 2019), which could be used to improve treatment options.

A U T H O R C O N T R I B U T I O N S
JAS and EO supervised the work described.SO, SI, JAS and EO made contributions to the design of the work.SO and SI contributed to the data collection.SO, SI, EO, MB-G, RN, IW, KW, and JAS contributed to the data analysis.SO, SI and JAS drafted the manuscript.All authors read, reviewed, and approved the final manuscript.

F
I G U R E 1 Family trio recruited and wider family pedigree affected by polycystic kidney disease.(a) The original trio recruited for WGS via Genomics England 100,000 Genomes Project.The shaded symbols indicate known cystic kidney disease.Arrow indicates proband.The proband's offspring are dizygotic twins, an affected daughter and a possibly affected son (marked as hashed shading) due to clinical doubt over cystic kidney phenotype.For the automated WGS analysis pipeline and filtering of variants by segregation the son was labelled as affected.(b) The wider family pedigree and confirmed clinical phenotypes of polycystic kidney disease.A review of the medical history, clinical notes and repeat imaging established a dominant pattern of inheritance and repeat phenotyping of the proband's son showed an absence of cystic kidney disease.The recruited trio are indicated via a red box.(c) Coronal and transverse abdominal MRI scans from the proband III:2 shows multiple bilateral kidney cysts, enlarged kidneys and a hepatic cyst (arrow).CT scans from the proband's father II:3 reveal multiple liver cysts (signified by arrows).II:3 had previously undergone bilateral nephrectomy for massive polycystic kidney disease.TA B L E 2 Classification of the variants found in the family.

F
I G U R E 2 Bespoke WGS analysis of the trio identifies a heterozygous ALG8 deletion and a heterozygous PKD1 pathogenic allele.(a) Visualization using Integrative Genomics Viewer (IGV) of the heterozygous ALG8 2509 bp deletion (Chr11:g.78105354_78107863del)which includes exon 10, in all members of the family trio III:2, IV:1 and IV:2.(b) PCR of genomic DNA to confirm ALG8 deletion and allow segregation of the variant from the proband's mother (II:4).A PCR product of 192 bp is only amplified in the presence of the 2509 bp ALG8 genomic deletion.The primer annealing sites and expected product sized are indicated above.(c) Visualization using IGV of the heterozygous PKD1 allele c.1889del; p.(Pro630Argfs*155) in the proband and female affected child.(d) Sanger sequencing of whole blood DNA from individual II:3 confirming the presence of the heterozygous PKD1 c.1889del; p.(Pro630Argfs*155) allele.(e) Updated wider family pedigree and confirmed clinical phenotypes of polycystic kidney disease, annotated with ALG8 and PKD1 genotypes following segregation testing.
Clinical phenotypes of family members.Age of onset, age at which kidney ultrasound identified kidney cysts.
TA B L E 1